IT support and especially IT innovations represent crucial success factors in investment management and financial decision making. The application of machine learning techniques for intraday decision support constitutes such a crucial factor, but existing literature has not been capable of melding latest research results from both the text and the capital market research field. Combining fundamental research results from both domains, it is possible to develop a forecasting model that performs better than both alternative internal (random) benchmarks and external prior research benchmarks. The approach is tested by means of “classical” evaluation methods such as well as by a simulation.